128GB max RAM · 4 bays · 2 x 2.5GbE
AINAS Score 40
No verified price

AI CapableTerraMaster
Why AI Capable: Intel N95 (4 cores, UHD Graphics, no NPU) with 32GB max DDR5 and no PCIe, USB4 or OCuLink GPU path, so only light CPU/iGPU inference via Docker.
Updated
Can the TerraMaster F8 SSD run local AI?
Partly. The TerraMaster F8 SSD runs local models on Intel N95 alone, and its 32GB ceiling allows models up to the 14B class, with 32B class models a tight fit (Estimated), but replies will be slow without graphics acceleration.
No graphics card can be fitted, which keeps every model on the CPU.
Local LLM support is documented by the manufacturer.
The F8 SSD suits a buyer who wants a compact all-flash NAS with a 10GbE port and treats AI as a light extra. It has 8 M.2 slots and no drive bays, so every disk is an NVMe SSD. The Intel N95 has 4 cores and Intel UHD Graphics, which handles hardware transcoding up to 4K60. It ships with 8GB of DDR5 and takes up to 32GB. TOS 7 runs Docker and virtual machines, and Ollama and Open WebUI are vendor confirmed as Docker containers. TOS 7 Photos adds AI face and scene recognition. Surveillance Manager handles up to 32 ONVIF cameras. The AINAS Score is 39, which fits a storage-first system. It carries a 2 year warranty.
The limit that rules it out for some buyers is the lack of any GPU path. There is no PCIe slot, no USB4, no Thunderbolt and no OCuLink, so you cannot add a graphics card later. Anyone wanting local LLM work beyond light use, such as private RAG, creative AI or larger models at speed, will hit that wall. The N95 has no NPU, so inference runs on the CPU and iGPU only. Model fit is a memory calculation, and Qwen3 32B fits at Q4 (Estimated), but that says nothing about how usable it is. Buyers in that group should look for a PCIe slot, OCuLink or USB4 for a GPU, or an NPU with a high TOPS rating, and memory above 32GB.
Amazon UK
8GB
£589.99
Verified
In stock
Buy on AmazonProduct prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon.co.uk or Amazon.com, as applicable, at the time of purchase will apply to the purchase of this product.
Each capability shows how it was confirmed: Verified by AINAS, Vendor confirmed, Community confirmed or Unverified.
| Capability | Supported | Evidence | Note |
|---|---|---|---|
| Local LLM | Limited | Vendor confirmed | CPU/iGPU-only inference via Docker; no NPU or GPU expansion. TerraMaster blog describes deploying Ollama/Open WebUI via Docker Manager on x86 TOS NAS generally (names F4-425 Plus and F4-424 Pro). |
| Ollama | Yes | Vendor confirmed | Runs as a Docker container via TOS Docker Manager per TerraMaster blog (x86 TOS NAS generally). |
| Open WebUI | Yes | Vendor confirmed | Open WebUI container via Docker; same TerraMaster blog source. |
| RAG | Limited | Unverified | Possible through Open WebUI or similar containers; no TerraMaster or community source showing RAG on TOS. |
| AI search | Limited | Vendor confirmed | TOS 7 OpenClaw agent and AI tagging allow natural-language queries; setup uses an API key from a provider such as OpenAI, so cloud LLM. Global search is an inverted index, not semantic. |
| GPU expansion | No | Vendor confirmed | Spec lists no PCIe slot; no USB4, Thunderbolt or OCuLink. |
| Hardware acceleration | Yes | Vendor confirmed | Hardware transcoding engine H.264/H.265/MPEG-4/VC-1 up to 4K60 via Intel iGPU. Video only, no NPU. |
| AI photos | Yes | Vendor confirmed | TOS 7 Photos: AI face recognition and AI scene recognition (80+ categories). |
| AI video | No | Unverified | No AI video feature found in TOS 7 material. |
| Surveillance AI | Yes | Vendor confirmed | TOS 7 Surveillance Manager supports up to 32 ONVIF IP cameras. |
Calculated from the F8 SSD's memory and each model's published size at each quantization. Rows marked Estimated are calculations, not benchmarks. Check any model.
| Model | Best quantization | Runs? | Memory needed | Expected performance | Verification |
|---|---|---|---|---|---|
| Llama 3.2 3BLlama, 3.21B | Q8 | Yes | 13GB | Not measured | Estimated |
| Llama 3.1 8BLlama, 8.03B | Q8 | Yes | 19GB | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | Q8 | Yes | 19GB | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | Q8 | Yes | 24GB | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | Q8 | Yes | 28GB | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | Q8 | Yes | 28GB | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | Q4 | Yes | 25GB | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | Q4 | Yes | 29GB | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | Q4 | Yes | 31GB | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | Q4 | Yes | 32GB | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | Q4 | Yes | 32GB | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | None fits | No | 60GB at Q4 | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | None fits | No | 89GB at Q4 | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | None fits | No | 65GB at Q4 | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | None fits | No | 96GB at Q4 | Not measured | Estimated |
One verdict per workload, from the F8 SSD's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The F8 SSD is a poor choice for a local LLM. Support is limited and vendor confirmed, with CPU and iGPU inference only through Docker, no NPU and no GPU expansion, though Ollama itself is confirmed as a container.
Private RAG is a poor fit here. RAG support is limited and Unverified, since it is only possible through Open WebUI or similar containers and no TerraMaster or community source shows it working on TOS.
AI search is poor on the F8 SSD. TOS 7 offers natural-language queries through the OpenClaw agent and AI tagging, but setup uses a cloud provider's API key, and global search is an inverted index, not semantic.
The verdict for AI photos is OK. TOS 7 Photos provides AI face recognition and AI scene recognition across 80+ categories, vendor confirmed, and the 8 M.2 slots give fast storage for a library, though the N95 has no NPU.
The F8 SSD is OK for surveillance. Surveillance Manager in TOS 7 supports up to 32 ONVIF IP cameras, vendor confirmed, and the Intel iGPU provides hardware transcoding for H.264 and H.265 video up to 4K60 (video only, no NPU).
The F8 SSD is OK for a home lab. TOS 7 runs Docker and virtual machines, and the N95 takes up to 32GB of non-ECC DDR5 SODIMM memory, with 8 M.2 slots and a 10GbE port for fast network access.
The verdict for business use is OK. It offers a 2-year warranty, TOS 7 with Docker and VM support, a 10GbE port and 8 M.2 slots for all-flash shared storage, but memory is non-ECC and there is only one network port.
Creative AI is a poor fit. There is no GPU expansion (vendor confirmed), because the spec lists no PCIe slot, USB4, Thunderbolt or OCuLink, and local LLM support is limited to CPU and iGPU inference, with hardware acceleration covering video transcoding only.
Out of the box the F8 SSD has 8GB; the official ceiling is 32GB of DDR5 non-ECC SODIMM. A memory upgrade is the cheapest way to make it more useful for AI.
Working backwards from 32GB: reserve about 8GB for the operating system and storage services, and at roughly 0.6GB per billion parameters at 4-bit, the largest model that fits is around 40B (Estimated).
There is no supported way to add a GPU to the F8 SSD: no PCIe slot and no OCuLink. Inference runs on the CPU, with Intel UHD Graphics handling media tasks, so keep to small models.
For a personal library, yes. A compact embedding model and a small chat model will run, but the F8 SSD is better as the document store behind a RAG system whose model runs on stronger hardware.
Built-in networking: 10GbE. Plenty for a shared knowledge base with several users.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 8GB | 8GB | None | Only configuration; 1x 8GB DDR5, 1 slot. |
The closest TerraMaster model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Business at 45 and Home Lab at 44, which points to small offices that need shared storage with private AI tools and room to grow.
Look elsewhere if you are buyers who expect to add a graphics card later.
Other AI Capable systems closest in price.
Buying guides that list it
Yes, on paper. DeepSeek-R1 32B (Qwen distill) fits at Q4 (Estimated), needing 20GB against a 32GB maximum, but it does not fit at Q8 (Estimated), which needs 35GB. The sold configuration has 8GB, so you would need a memory upgrade first. This is a memory calculation, not a benchmark, and inference would run on the CPU and iGPU only because there is no NPU or GPU path.
Yes. The F8 SSD ships with 8GB of DDR5 non-ECC SODIMM memory in one slot and takes up to 32GB. The memory is expandable rather than soldered. It does not support ECC, so choose non-ECC DDR5 SODIMM modules. The only configuration sold is the 8GB one, so any upgrade is something you add yourself.
No. The F8 SSD has no PCIe slot, and no USB4, Thunderbolt or OCuLink ports, so there is no way to connect a graphics card. Vendor documentation confirms GPU expansion is not supported. Graphics work is limited to the Intel UHD Graphics inside the N95, which provides video hardware transcoding up to 4K60. It is video only, with no NPU.
Yes. The F8 SSD has 8 M.2 slots and 0 drive bays, so every drive is an NVMe SSD and there is no room for hard disks. The 8GB configuration is the only one, and all eight slots share a single 10GbE port for network access. Memory sits in one DDR5 slot, with 8GB fitted and a 32GB maximum, so the flash storage and the memory ceiling are separate limits to plan around.
Yes. TOS 7 runs Docker through Docker Manager, and TerraMaster describes Ollama and Open WebUI running as containers on its x86 NAS systems. Both are vendor confirmed. Local LLM support is rated limited, because inference is CPU and iGPU only with no NPU or GPU expansion. Private RAG is only described as possible and remains Unverified. TOS 7 also supports virtual machines.
Yes. Ollama support is documented by the manufacturer. See the capability table above for the source.
Yes. Open WebUI support is documented by the manufacturer. See the capability table above for the source.
Yes. AI photos support is documented by the manufacturer. See the capability table above for the source.
Yes. Surveillance AI support is documented by the manufacturer. See the capability table above for the source.
Yes. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has 10GbE built in. The fastest port is 10GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.